Xavier Talwatte

Case Studies

About

Xavier Talwatte

Uber Eats Feature Design

Project Brief

Uber Eats enables global users to order delivery from local restaurants conveniently. However, users seeking to learn more about restaurants with few ratings often leave the app in search of more detailed information. Our aim was to empower users to make informed decisions when selecting a restaurant for delivery.

My Contributions

I took ownership of:

Creating low-fidelity, mid-fidelity, and high-fidelity prototypes
Conducting user interviews and testing
Project management
Leading ideation activities
Initial concept

Results

The project resulted in improved user confidence in food safety and higher satisfaction with the restaurant discovery experience. This led to increased likeliness of ordering from a restaurant. While further research is needed, increased satisfaction with restaurant information could correlate with lower drop-off rates.

Team (5 UX Designers)

Role

UX Designer

Timeline

3-week design sprint + 1 week additional work

+20%

Confidence in Food Safety

+16%

Conversion Rate

+36%

Satisfaction With Restaurant Details

Uber Eats Feature Design

Project Brief

My Contributions

I took ownership of:

Creating low-fidelity, mid-fidelity, and high-fidelity prototypes
Conducting user interviews and testing
Project management
Leading ideation activities
Initial concept

Results

Team (5 UX Designers)

Role

UX Designer

Timeline

3-week design sprint + 1 week additional work

+20%

Confidence in Food Safety

+16%

Conversion Rate

+36%

Satisfaction With Restaurant Details

Uber Eats Feature Design

Project Brief

My Contributions

I took ownership of:

Creating low-fidelity, mid-fidelity, and high-fidelity prototypes
Conducting user interviews and testing
Project management
Leading ideation activities
Initial concept

Results

Team (5 UX Designers)

Role

UX Designer

Timeline

3-week design sprint + 1 week additional work

+20%

Confidence in Food Safety

+16%

Conversion Rate

+36%

Satisfaction With Restaurant Details

Team (5 UX Designers)

Role

UX Designer

Timeline

3-week design sprint + 1 week additional work

Project Brief

My Contributions

I took ownership of:

Creating lofi, midfi, and hifi prototypes
Conducting user interviews and testing
Project management
Leading ideation activities
Initial concept

My Contributions

I took ownership of:

Creating low-fidelity, mid-fidelity, and high-fidelity prototypes
Conducting user interviews and testing
Project management
Leading ideation activities
Initial concept

Results

+20%

Confidence in Food Safety

+16%

Conversion Rate

+36%

Satisfaction With Restaurant Details

Process: Discover, Define, Develop, Deliver

Discover:

This project was initially inspired by a burrito run to one of my favorite local Mexican restaurants. I was going to order delivery but instead opted to pick up the food myself. When I arrived, I saw a sign in the window that made me lose my appetite.

A concerning realization for me was that had I decided to order online, I wouldn't have known that the restaurant had failed a health inspection. This got me thinking, "What are other discrepancies between online and in-person ordering processes that inhibit users' ability to have confidence in the safety and quality of their food?"

Initial Research:

Our team surveyed 18 participants and interviewed 6 participants to learn about users' ordering processes, the extent to which they value health information, and what they look for when deciding where to order from.

Survey

Over half of the participants experienced a foodborne illness after dining at a restaurant or ordering takeout.

One in three participants was aware of food safety ratings in place in their area.

Over half of participants had an experience where food safety information influenced their decision to order from a restaurant.

User Interviews

Survey

Over half of participants experienced a foodborne illness after dining at a restaurant or ordering takeout.

One in three participants was aware of food safety ratings in place in their area.

Over half of participants had an experience where food safety information influenced their decision to order from a restaurant.

Define:

Using data from the survey and interviews as a starting point, we refined our understanding of the problem by creating affinity diagrams and feature prioritization matrices.

Affinity Diagram

Rather than having you squint to read hundreds of tiny post-its, here are the grouping categories we used.

Feature Prioritization Matrix

We prioritized frequently occurring wants, needs, and pain-points with a high impact-to-complexity ratio.

Problem statement:

Uber Eats users ordering from a new restaurant for the first time need a convenient way to learn more about the quality and safety of food so that they can order with confidence.

We conducted a pre-test to establish some baseline KPIs and observe how users currently use the app.

5 participants
Remote, moderated tests
User satisfaction ratings

3 out of 5 participants left the Uber Eats app to look for additional information in Yelp or Google Maps.

Key findings:

Users who looked at alternative sources (such as Yelp and Google Maps) were 23% more confident in both the safety and quality of food.
Users who consulted external sources were also 10% more likely to place an order.
Users rely heavily on images of food, written reviews, and star ratings to determine food safety and quality.

"I can't click into this to read more about the reviews, so I don't know what their rating system is. I don't know what the metrics they're using are, and I just feel like it's really unreliable."

Develop:

During this stage of the process, I contributed through:

Crazy Eights brainstorming sketches
Defining user flows
Creating low-fidelity, mid-fidelity, and high-fidelity wireframes and prototypes
Conducted multiple rounds of user testing
Synthesizing user feedback and iterating based on findings

Here's a glance at some of the things I worked on:

Crazy Eights Brainstorm Sketches

In eight minutes, I created eight sketches for potential solutions in a group brainstorming session. We each presented our ideas and starred features from each other's designs to incorporate into initial wireframes.

Features from my sketches that were selected included:

The option to view user-written reviews by clicking on ratings. (Multiple users tried to do this during pre-testing.)
A review modal that allows users to quickly review restaurants they've recently ordered from for an incentive.

Creating Low-Fidelity Versions of Existing Screens

I selected initial screenshots from the Uber Eats app to incorporate into our user flow based on screens users visited during the pre-test (personal details obscured or altered for privacy).

Using initial screenshots and Uber's Base design system as reference points, I made pixel-perfect low-fidelity wireframes to make it easy to modify existing designs to accommodate changes. This process also helped define visual design guidelines for the new screens we created.

Creating New Screens for the Low-Fidelity Prototype

Using ideas from group brainstorming sessions, I created four new screens and altered one existing screen (two screens created with assistance from a team member).

Restaurant Details: Modified to allow users to view rating details and include a health score category based on local health rating systems.

Review Page: Showcase user reviews within the Uber Eats app to decrease drop-off rates due to visiting competing sites with reviews.

Photo Page: Allows users to view user-uploaded food photos from a restaurant.

Review Modal: Users can leave reviews of restaurants they've ordered from the next time they open the app. Features a star rating, individual menu item rating, and written review.

Confirmation Modal: Confirmation of submission of user's review and application of promotion. Provides a preview of what the review will look like to others.

Testing the Low-Fidelity Prototype

Clarity

3/5 users thought the Health Score needed clarification.

Clarity

3/5 users thought the Health Score needed clarification.

Navigation

4/5 participants struggled to find the restaurant details page.

Navigation

4/5 participants struggled to find the restaurant details page.

Visibility

2/5 participants didn’t notice the promotional offer.

Concept Validation

All participants liked the idea of the review modal.

During the initial round of usability testing, we found several areas of confusion, as well as indications that we were moving in the right direction. I contributed to revisions by:

Creating a detailed health-score information page
Increasing the size of hotspots to reduce error rates
Changing microcopy hierarchy and content to emphasize the promotional offer

Visibility

2/5 participants didn’t notice the promotional offer.

Concept Validation

All participants liked the idea of the review modal.

I used existing detail screens in Uber Eats as a basis for the designs of the Health Score Details page, including creating icons that mimic the style of Uber Base icons. The design also utilizes a sample source (Hazel Analytics, a company that partners with competitor Yelp to provide this information) to demonstrate how information could be collected.

Create Mid-fidelity Wireframes

After incorporating feedback from the initial round of testing, I developed a mid-fidelity prototype and conducted another round of testing to assess whether the iterations addressed the issues identified by users.

Mid-Fidelity Testing and Iterations

The additional round of testing revealed that users who didn't have a strong feeling about a restaurant were less likely to leave a review.

To make the review-writing process more convenient, I added a new section to the review page that allows users to select from pre-populated shoutouts.

This resulted in many more users expressing a desire to leave reviews.

Deliver:

Since this started as a class project that was initially volunteer work, our final goal was to convince stakeholders to invest in making the project a reality. I created a presentation to communicate the process, reasoning, and outcomes with key stakeholders, which ultimately led to stakeholder buy-in and implementation of our work (development currently in progress).

After securing stakeholder buy-in, we met with the developer to evaluate the feasibility of our designs given constraints such as:

CMS technical constraints
Budget and timeframe constraints
Organizational constraints
Content-related constraints

After meeting with the developer, I worked with the team to identify key revisions to the style guide to clarify documentation about specific use cases for certain components, guidelines, and edge cases. Though I am not able to continue working on preparation for development due to schedule conflicts, I helped the team outline the next steps for developer handoff.

Key Features from Final Designs:

Health score information allows users to quickly build trust in the safety of a restaurant's food.

Browsing user-written reviews allows users to see what others are saying about a restaurant so they can make informed decisions without leaving the app.

Leaving user-written reviews helps restaurants build reputations through shoutouts and empowers users to share their experiences.

Conclusion

Since this was a conceptual project, there wasn't a pressing need to create deliverables for handoff outside of the ones we used internally for organization. However, should this project be taken up in the future, I would revisit, clarify and provide documentation for the following areas:

Interaction details for when users click on images uploaded by other customers
Specifics regarding the user journey of leaving a review (timing, notifications, etc.)
Process for uploading user-generated images (our current design does not account this)

What I learned:

Importance of having a clearly defined user flow: Spending more time on establishing the user flow would have helped me identify all of the necessary screens from the original Uber Eats app. I would have wasted less time recreating unmodified screens from scratch.
Project management: Being in charge of project management (through Trello) helped our team establish transparent communication about deadlines for project milestones in a remote work environment.
Consider multiple user groups: Upon completing the project, I recognized that we only interviewed one group of users. We did not speak with restaurants using Uber Eats to understand how this feature would affect their business. To address this, I spoke with four managers at local restaurants in my area and received overwhelmingly positive initial support. However, I did not have the opportunity to speak with restaurants that received low health scores, and future iterations of this project should include these perspectives.

Interaction details for when users click on images uploaded by other customers
Specifics regarding the user journey of leaving a review (timing, notifications, etc.)
Process for uploading user-generated images (our current design does not account this)

What I learned:

Importance of having a clearly defined user flow: Spending more time on establishing the user flow would have helped me identify all of the necessary screens from the original Uber Eats app. I would have wasted less time recreating unmodified screens from scratch.
Project management: Being in charge of project management (through Trello) helped our team establish transparent communication about deadlines for project milestones in a remote work environment.
Consider multiple user groups: Upon completing the project, I recognized that we only interviewed one group of users. We did not speak with restaurants using Uber Eats to understand how this feature would affect their business. To address this, I spoke with four managers at local restaurants in my area and received overwhelmingly positive initial support. However, I did not have the opportunity to speak with restaurants that received low health scores, and future iterations of this project should include these perspectives.

Thank you for taking the time to read this case study!

+20%

Confidence in Food Safety

+16%

Conversion Rate

+36%

Satisfaction With Restaurant Details

Xavier Talwatte | UX Designer